The complex chemistry of peroxynitrite decomposition: New insights

Article abstract of DOI:10.1021/ja980871x

The yield of hydroxyl radicals produced in the decomposition of peroxynitrous acid (HOONO) at room temperature in deoxygenated and bicarbonate free water at pH ~6.8 has been determined to be roughly 10%. This value rests on a detailed study of the decomposition of peroxynitrous acid in the presence of dimethyl sulfoxide with stopped-flow kinetics and product analyses by unequivocal methods. The HO·/DMSO reaction is known to yield methane sulfinic acid (MSA) and CH₃· radicals with 91% efficiency. MSA was quantified by ¹H NMR and its measured yield was corrected to allow for its extensive further oxidation to methane sulfonic acid. Methyl radicals were quantified by trapping with a water-soluble, stable nitroxide. At low peroxynitrite concentrations these two techniques gave HO· yields of ca. 8% and ca. 13%, respectively. We conclude that in water the main (ca. 90%) decomposition pathway for peroxynitrite involves a rearrangement to nitric acid via an in-cage collapse of the singlet HO·/·NO₂ radical pair which may, in part, be preceded by electron transfer to form an HO^-/⁺NO₂ intimate ion pair. We emphasize that, in contrast to many earlier reports, a distinct pathway to hydroxyl radicals is present, which implies that a significant portion of the oxidative nature of peroxynitrite can stem from hydroxyl radical-induced chemistry.

Full text of DOI:10.1021/ja980871x

J. Am. Chem. Soc. 1998, 120, 7211-7219
7211
The Complex Chemistry of Peroxynitrite Decomposition: New
Insights1
C. E. Richeson,2 P. Mulder,*,3 V. W. Bowry,4 and K. U. Ingold*
Contribution from the Steacie Institute for Molecular Sciences, National Research Council of Canada,
Ottawa, Ontario, Canada K1A 0R6
ReceiVed March 16, 1998
Abstract: The yield of hydroxyl radicals produced in the decomposition of peroxynitrous acid (HOONO) at
room temperature in deoxygenated and bicarbonate free water at pH ∼6.8 has been determined to be roughly
10%. This value rests on a detailed study of the decomposition of peroxynitrous acid in the presence of
dimethyl sulfoxide with stopped-flow kinetics and product analyses by unequivocal methods. The HO•/DMSO
•
reaction is known to yield methane sulfinic acid (MSA) and CH3 radicals with 91% efficiency. MSA was
1
quantified by H NMR and its measured yield was corrected to allow for its extensive further oxidation to
methane sulfonic acid. Methyl radicals were quantified by trapping with a water-soluble, stable nitroxide. At
low peroxynitrite concentrations these two techniques gave HO• yields of ca. 8% and ca. 13%, respectively.
We conclude that in water the main (ca. 90%) decomposition pathway for peroxynitrite involves a rearrangement
to nitric acid via an in-cage collapse of the singlet HO•/•NO2 radical pair which may, in part, be preceded by
electron transfer to form an HO-/+NO2 intimate ion pair. We emphasize that, in contrast to many earlier
reports, a distinct pathway to hydroxyl radicals is present, which implies that a significant portion of the
oxidative nature of peroxynitrite can stem from hydroxyl radical-induced chemistry.
+
H
In a 1990 landmark publication, Beckman et al.5 suggested
that two relatively unreactive, but biologically important, free
-OONdO y-H+z HOONdO f HO• + NO2
(2)
•
radicals, superoxide and nitric oxide, would combine under
physiological conditions to form peroxynitrite.6 Reaction 1
Beckman et al.’s suggestion that peroxynitrite could undergo
O-O bond homolysis, reaction 2, has had a difficult time
because it has proved impossible to achieve consensus regarding
the yield of freely diffusing HO• radicals relative to the yield
of the isomerization, or in-cage collapse, product, nitric acid.
appears to be diffusion controlled (k1 ) 1.9 × 1010 M-1 s-1 7
)
O2•- + •NO f -OONdO
(1)
kf
•
H+ + NO3- r HOONO y
\
z [HO• + NO2 ]cage f
and the anion is thermally stable. However, at physiological
pH, the anion will be partially protonated (pKa 6.8)6-8 to form
peroxynitrous acid, which has a half-life of just less than 1 s at
37 °C. Furthermore, Beckman et al.5 reported that peroxynitrite
decomposition9 generated “a strong oxidant able to initiate many
reactions currently used to implicate the action of HO•”. These
workers went on to make the exciting suggestion that this strong
oxidant might actually be the hydroxyl radical (which is by far
the most reactive of all biologically relevant “oxy radicals”).
kr
H+ + NO3 (3)
-
Thus, Beckman et al.5 reported maximum “product yields (at
infinite detector concentration relative to peroxynitrite) indicative
of hydroxyl radical” at pH 6.0 of 7.0 ( 0.1% (based on
malondialdehyde, MDA, produced by the oxidation of deox-
yribose) and 24 ( 1.0% (based on formaldehyde produced by
the oxidation of dimethyl sulfoxide, DMSO). The apparent
yields of free HO• radicals (or of a hydroxyl radical-like species)
based both on MDA and formaldehyde production decreased
at higher pH and there are conflicting explanations for these
observations.10,11 The yield of hydroxyl radicals (or, sometimes,
of a hydroxyl radical-like species) from the decomposition of
peroxynitrite has also been reported to be 40%,10 32%,12 25%,11
4%,13 about 1%,14 0.9%13 and zero15 or probably zero.7,16-19
(1) Issued as N.R.C.C. No. 40880.
(2) N.R.C.C. Summer student 1994 and 1995.
(3) Gorlaeus Laboratories, Leiden University, 2300 RA Leiden, The
Netherlands.
(4) Department of Chemistry, University of New England, Armidale,
N.S.W. 2351, Australia.
(5) Beckman, J. S.; Beckman, T. W.; Chen, J.; Marshall, P. A.; Freeman,
B. A. Proc. Natl. Acad. Sci. U.S.A. 1990, 87, 1620-1624.
(6) For a wonderfully comprehensive and critical review of the chemistry
of peroxynitrite, see: Pryor, W. A.; Squadrito, G. L. Am. J. Physiol. 1995,
268, L699-L722.
(10) Yang, G.; Candy, T. E. G.; Boaro, M.; Wilkin, H. E.; Jones, P.;
Zazhat, N. B.; Saadalla-Nazhat, R. A.; Blake, D. R. Free Radical Biol.
Med. 1992, 12, 327-330.
(7) Kissner, R.; Nauser, T.; Bugnon, P.; Lye, P. G.; Koppenol, W. H.
Chem. Res. Toxicol. 1997, 10, 1285-1292.
(11) Crow, J. P.; Spruell, C.; Chen, J.; Gunn, C.; Ischiropoulos, H.; Tsai,
M.; Smith, C. D.; Radi, R.; Koppenol, W. H.; Beckman, J. S. Free Radical
Biol. Med. 1994, 16, 331-338.
(8) Pryor, W. A.; Jin, X.; Squadrito, G. L. Proc. Natl. Acad. Sci. U.S.A.
1994, 91, 11173-11177.
(12) Mahoney, L. R. J. Am. Chem. Soc. 1970, 92, 5262-5263.
(13) Pou, S.; Nguyen, S. Y.; Gladwell, T.; Rosen, G. M. Biochim.
Biophys. Acta 1995, 1244, 62-68.
(9) Peroxynitrite will be used henceforth to refer to the sum of -OONO
and HOONO.6.
S0002-7863(98)00871-3 CCC: $15.00 Published 1998 by the American Chemical Society
Published on Web 07/11/1998

7212 J. Am. Chem. Soc., Vol. 120, No. 29, 1998
Richeson et al.
However, many of these reports were published before the
important role that CO2 can play in peroxynitrite chemistry was
known.20,21 Adventitious CO2 could have perturbed some of
these measurements of HO• yields.22 Furthermore, some of the
substrates used to measure the HO• yields probably reacted
directly with peroxynitrite. Additional uncertainties regarding
the yield of HO• radicals from peroxynitrite come from the
variable effects of known hydroxyl radical scavengers on the
yields of oxidation products from various substrates.5,6,8,11,12,23
The peroxynitrite story became even more confusing when
Koppenol et al.24 claimed that their thermodynamic calculations
and kinetic measurements precluded the formation of HO• from
HOONO. The kinetic argument was based on the measured
activation parameters for peroxynitrite decomposition, viz.,24
∆Hq ) 18 ( 1 kcal mol-1 and ∆Sq ) 3 ( 2 cal mol-1 K-1
(corresponding to log(A/s-1) ) 13.9). Although the calculated
activation enthalpy for HOONO homolysis in water, viz.,24 17
kcal mol-1, was in good agreement with the measured value,
homolysis was discarded because the activation entropy was
smaller than the ca. 12 cal mol-1 K-1 generally found for
homolysis of the O-O bond in peroxides. The small magnitude
of the activation entropy for peroxynitrite decomposition relative
to the homolysis of other peroxides must indeed reflect “rigidity
in the transition state” (or extensive in-cage return,6 kr, vide
infra) if the former reaction is homolytic. However, the potential
role of the water solvent in producing such “rigidity” appears
not to have been considered until some years later.6 Subsequent
thermodynamic and kinetic arguments have both supported25
and opposed26 Koppenol et al.’s original conclusion that
HOONO does not undergo homolysis in water. Most recently,
Koppenol and Kissner27 have admitted that homolysis is possible
but, nevertheless, conclude that homolysis is unlikely.
Measurements of the rates of decay of peroxynitrite in the
presence of, for example, thiols have shown that the reaction is
first order in peroxynitrite and first order in thiol, which is
consistent with a simple bimolecular reaction.24,28 There are
quite a number of other “substrates”, however, which have been
reported to react with kinetics that are first order in peroxynitrite
but zero order in “substrate”.6 That is, the rate of decay of
peroxynitrite is the same in the absence as in the presence of
these substrates. (The occurrence of such reactions can, of
course, only be revealed by product analyses.) One consequence
of the thermochemical theories extant at the time was that these
zero-order in substrate reactions were assumed to involve some
high energy form of peroxynitrite that was not the HO• radical.
This high-energy form has been designated as HOONO* 6 and
has often been assumed to be present together with ground-
state HOONO in peroxynitrite solutions. Initially, the HOONO*
was a “vibrationally excited” form of HOONO11 but, as Pryor
and Squadrito6 have correctly pointed out, “vibrationally excited
states generally are too short-lived (with lifetimes of ∼10-11 s)
to participate in bimolecular reactions”. As an alternative to
the kinetically incompetent vibrationally excited state, Pryor and
co-workers6,8,16,19,29 have suggested that HOONO* is a high-
energy, metastable form of HOONO that is present in steady
state with ground-state peroxynitrous acid.
The foregoing is intended to provide some insight into the
complexities of peroxynitrite chemistry and to set the stage for
our own work. We limited our objective to determining to what
extent, if at all, hydroxyl radicals are formed during the thermal
decomposition of peroxynitrite at room temperature.
Results
Our work has been based on the following concepts: (1) use
of a substrate that (i) is known to react with HO• radicals at, or
close to, the diffusion-controlled rate, (ii) is at least purported
not to react with peroxynitrite, and (iii) is known to give a clearly
defined product (or products) in its reaction with HO•; (2) study
of the kinetics of the reactions (if any) of peroxynitrite with (i)
the initial substrate, (ii) its HO•-derived product(s), and (iii) any
secondary reagent added to “trap” any HO•-derived product;
and (3) use of unequivocal analytical procedures to search for
the “expected ” HO•-derived product(s) in a manner that would
allow their identification and quantification even if formed in a
very low yield from the peroxynitrite.
(14) van der Vliet, A.; O’Neill, C. A.; Halliwell, B.; Cross, C. E.; Kauer,
H. FEBS Lett. 1994, 339, 89-92.
(15) Shi, X.; Lenhart, A.; Mao, Y. Biochem. Biophys. Res. Commun.
1994, 203, 1515-1521.
(16) Lemercier, J.-N.; Squadrito, G. L.; Pryor, W. A. Arch. Biochem.
Biophys. 1995, 321, 31-39.
(17) Dikalov, S.; Kirilyuk, I.; Grigor’ev, I. Biochem. Biophys. Res.
Commun. 1996, 218, 616-622.
(18) Narayan, M.; Berliner, L. J.; Merola, A. J.; Diaz, P. T.; Clanton, T.
L. Free Radical Res. 1997, 27, 63-72.
(19) Pryor, W. A.; Jin, X.; Squadrito, G. L. J. Am. Chem. Soc. 1996,
118, 3125-3128.
(20) Lymar, S. V.; Hurst, J. K. J. Am. Chem. Soc. 1995, 117, 8867-
8868.
(21) Uppu, R. M.; Squadrito, G. L.; Pryor, W. A. Arch. Biochem. Biophys.
1996, 327, 335-343. Uppu, R. M.; Pryor, W. A. Biochem. Biophys. Res.
Commun. 1996, 229, 764-769. Denicola, A.; Freeman, B. A.; Trujillo, M.;
Radi. R. Arch. Biochem. Biophys. 1996, 333, 49-58. Zhang, H.; Squadrito,
G. L.; Uppu, R. M.; Lemercier, J.-N.; Cueto, R.; Pryor, W. A. Arch.
Biochem. Biophys. 1997, 339, 183-189. Pryor, W. A.; Lemercier, J.-N.;
Zhang, H.; Uppu, R. M.; Squadrito, G. L. Free Rad. Biol. Med. 1997, 23,
331-338. Lemercier, J.-N.; Padmaja, S.; Cuerto, R.; Squadrito, G. L.; Uppu,
R. M.; Pryor, W. A. Arch. Biochem. Biophys. 1997, 345, 160-170.
(22) We did extensive earlier work on HO• yields from peroxynitrite
(which was written up and submitted!) before we appreciated (thanks to
Prof. W. A. Pryor) the important role that CO2 can play in peroxynitrite
chemistry.20,21 No precautions were taken in our earlier work to exclude
CO2 and although hydroxyl radicals were detected (using some of the
methodology described herein) their yield was very much lower (ca. 0.4%).
This low yield was probably due partly to the presence of CO2 and partly
to the fact that significantly higher peroxynitrite concentrations were used
in the earlier work (vide infra).
Experimental Approach. Following in Beckman et al.’s
footsteps,5 we chose DMSO as the substrate to probe for HO•
radicals since this compound was not supposed to react with
peroxynitrite.5,11 (There is probably a direct reaction between
peroxynitrite and DMSO but it is extremely slow, vide infra.)
In water at ambient temperatures DMSO reacts extremely
rapidly with hydroxyl radicals (k4 ) 7 × 109 M-1 s-1) to yield
methane sulfinic acid (MSA) and methyl radicals with a ca.
91% efficiency.30 In the event that peroxynitrite did yield
•
•
(CH ) SO CH S(O)OH
HO +
f
+ CH3
(4)
3 2
3
DMSO
MSA
hydroxyl radicals, an additional attraction of DMSO was the
formation of two products which should permit two independent
measurements of the HO• yield.
(23) For a detailed listing of scavenger effects see: Goldstein, S.;
Squadrito, G. L.; Pryor, W. A.; Czapski, G. Free Radical Biol. Med. 1996,
21, 965-974.
(24) Koppenol, W. H.; Moreno, J. J.; Pryor, W. A.; Ischiropoulos, H.;
Beckman, J. S. Chem. Res. Toxicol. 1992, 5, 834-842.
(25) Bartlett, D.; Church, D. F.; Bounds, P. L.; Koppenol, W. H. Free
Radical Biol. Med. 1995, 18, 85-92.
(28) Radi, R.; Beckman, J. S.; Bush, K. M.; Freeman, B. A. J. Biol.
Chem. 1991, 266, 4244-4250.
(29) Squadrito, G. L.; Jin, X.; Pryor, W. A. Arch. Biochem. Biophys.
1995, 322, 54-59.
(26) Mere´nyi, G.; Lind, J. Chem. Res. Toxicol. 1997, 10, 1216-1220.
(27) Koppenol, W. H.; Kissner, R. Chem. Res. Toxicol. 1998, 11, 87-
89.
(30) Veltwisch, D.; Janata, E.; Asmus, K.-D. J. Chem. Soc., Perkin Trans.
2 1980, 146-153.

The Complex Chemistry of Peroxynitrite Decomposition
J. Am. Chem. Soc., Vol. 120, No. 29, 1998 7213
Peroxynitrite (e75 mM) was synthesized at pH 12 by reaction
of ozone with sodium azide following the procedure of Gleu
and Roell31 and of Pryor et al.32 with rigorous exclusion of
carbon dioxide20-22,33 (see Experimental Section).
peroxynitrite, which is first order in peroxynitrite and first order
in DMSO, because the experimental first-order rate constant
for peroxynitrite decay, kexptl, increases as the DMSO concentra-
tion is increased (see Supporting Information). We believe that
the most likely product from such a direct, bimolecular reaction
would be methyl sulfone, reaction 6, and this compound was
The kinetics of peroxynitrite decay were monitored at 302
nm at room temperature (23 ( 1 °C) in a stopped-flow
apparatus.5,8,10,16,19,24,25,28,29 The nitrogen-purged (deoxygenated
and CO2 free) pH 12 peroxynitrite solution was very rapidly
mixed (∼1.3 ms) with known concentrations of substrate in
nitrogen-purged (deoxygenated and CO2 free) KH2PO4 buffer
so that all subsequent reactions occurred at pH ∼6.8. (The pH
was measured after every run.) In the absence of substrate,
numerous measurements of the first-order rate constant for (2.6
mM) peroxynitrite decay, ko, gave an average value of 0.95 (
0.1 s-1 at room temperature. At higher peroxynitrite concentra-
tions the decay appeared to be faster.34 However, most of our
product studies, and the only ones we consider to be reasonably
reliable, were deliberately carried out with the minimum
concentration of peroxynitrite (2.6 mM), which gave accurately
measurable absolute yields of products by the techniques we
employed. Even at 10.5 mM peroxynitrite the calculated HO•
yields were dramatically lower than at 2.6 mM, vide infra. (The
dependence of HO• yields on peroxynitrite concentrations both
higher than 10.5 mM and lower than 2.6 mM is currently under
investigation.)
Me2SO + -OONO (HOONO) f
Me2SO2 + NO2- + (H+) (6)
indeed observed by NMR product analysis (vide infra). If the
kinetic data are plotted in the usual way according to kexptl ) ko
+ k6[DMSO], where k0 ()0.95 s-1, vide supra) is the rate
constant in the absence of DMSO, the derived value of k6 is
2.1 M-1 s-1
.
(2) MSA. This acid was found to react fairly rapidly with
peroxynitrite with kinetics that were first order in peroxynitrite
and first order in MSA (see Supporting Information). We
believe the most likely product of this reaction would be
methane sulfonic acid (reaction 7) and this compound was also
MeS(O)OH + -OONO (HOONO) f
MeS(O)2OH + NO2- + (H+) (7)
Products were generated from DMSO under similar condi-
tions in a separate rapid mixing (ca. 4 ms) apparatus with a
glass collection vessel attached to the mixer. These products
observed by NMR product analysis (vide infra). With use of
the equation kexptl ) k0 + k7[MSA], with ko ) 0.95 s-1, the
kinetic data yield k7 ) 325 M-1 s-1
.
1
were analyzed by H NMR.35 Methyl radicals were trapped
(3) TEMPOL. Rapid mixing of solutions to yield final
concentrations of peroxynitrite of 10.5 mM and TEMPOL
ranging from 5 to 25 mM showed that the decomposition of
peroxynitrite was accelerated by the nitroxide (reaction 8), see
by 4-hydroxy-2,2,6,6-tetramethylpiperidin-N-oxyl, TEMPOL,
which was added to the DMSO/buffer solution. This nitroxide
is very soluble (>1 M) in the buffer, and the resultant
O-methylhydroxylamine, CH3-TEMPOL, was analyzed by
GCMS with CD3-TEMPOL (added postreaction) as an internal
standard.
TEMPOL + -OONO (HOONO) f products
Supporting Information. According to kexptl ) ko + k8-
(8)
[TEMPOL] the magnitude of k8 was 44.5 M-1 s-1
.
Products. (1) 1H NMR Product Analyses. Aliquots of the
products formed by rapid mixing of peroxynitrite (concentrations
2.6 and 10.5 mM after mixing) and DMSO (concentration 105
1
mM after mixing) were analyzed by H NMR (400 MHz). A
typical spectrum is shown in Figure 1. In all samples four
products could be detected: methyl sulfone, Me2SO2; methane
sulfinic acid, MeS(O)OH (MSA); methane sulfonic acid,
MeS(O)2OH; and methanol (see Table 1). Unfortunately, the
1H NMR peak due to MeS(O)2OH was not resolved from the
peak due to DMSO at a DMSO concentration of 420 mM.
Formaldehyde, a product expected on the basis of earlier
work,5,32 was not detected. However, blank experiments with
[CH2O] ) [CH3OH] ) 0.15 mM in the buffer showed that
formaldehyde could not be detected by 1H NMR at this
concentration.
Kinetics. (1) DMSO. In contrast to earlier reports, there
would appear to be a slow reaction between DMSO and
(31) Gleu, K.; Roell, E. Z. Anorg. Allg. Chem. 1929, 179, 233-266.
(32) Pryor, W. A.; Cueto, R.; Jin, X.; Koppenol, W. H.; Ngu-Schwemlein,
M.; Squadrito, G. L.; Uppu, P. L.; Uppu, R. M. Free Radical Biol. Med.
1995, 18, 75-83.
(33) A referee of our earlier work pointed out that the ozone/azide route
to peroxynitrite might not have been the best synthetic choice because nitrite
is formed in only slightly lower concentration than peroxynitrite32 and nitrite
reacts with HO• radicals about as rapidly (see, e.g.: Logager, T.; Sehested,
K. J. Phys. Chem. 1993, 97, 6664-6669) as does DMSO.30 Thus, some
HO• would be “lost” by reaction with nitrite. However, on repeating all
our earlier work under CO2-free conditions we opted for the simplicity of
the ozone/azide method which can be justified by the fact that the DMSO
concentration was always 10, or more, times greater than the concentration
of peroxynitrite and, hence, than the concentration of nitrite.
(34) After this work was completed, ref 7 appeared, which reported that
the decay of peroxynitrite was slower at 0.48 mM than at 0.048 mM
[peroxynitrite]. Both of these concentrations are, unfortunately, below the
minimum peroxynitrite concentration we felt we could employ for reliable
quantitation of products.
Interestingly, the yields of HO• radical-derived products
decreased dramatically relative to the initial peroxynitrite
concentration as the latter was increased from 2.6 to 10.5 mM.
We attribute this to other (higher order) reactions of peroxynitrite
which appear to produce kinetic anomalies at higher peroxyni-
trite concentrations, vide supra,34 and presumably reduce the
fraction of peroxynitrite decaying by simple O-O bond
homolysis.
The apparent yields of methyl sulfone in these peroxynitrite/
DMSO reactions had to be corrected for the fact that this
compound was present as a very minor impurity [(8 × 10-2)%]
in the starting DMSO. Thus, in a typical experiment with post-
mixing [DMSO] ) 105 mM and [peroxynitrite] ) 2.6 mM,
(35) The popular analyses for MSA by its color-making reactions with
diazonium salts such as Fast Garnet GBC36 and Fast Blue BB37 did not
work in our DMSO/HOONO product mixtures, even when MSA was
deliberately added! The culprit is residual azide (used in the synthesis of
peroxynitrite), which destroys the diazonium salts even at very low
concentrations.

7214 J. Am. Chem. Soc., Vol. 120, No. 29, 1998
Richeson et al.
Figure 1. Typical 1H NMR spectrum (400 MHz) of the products of reaction of 2.6 mM peroxynitrite with 105 mM DMSO. The peaks (ppm) are
due to CH3OH (3.24), Me2SO2 (3.03), MeSO3H (2.70), DMSO (2.61) with two 13C satellites bands), MeSO2H (2.19), and an unknown impurity in
the DMSO (2.12).
1
Table 1. Product Yields Measured by H NMR (400 MHz) for
the Decomposition of Peroxynitrite at pH 6.8 and 23 ( 1 ˚C in the
Presence of DMSO
reason for this discrepancy is that the measured kinetics of the
reaction between peroxynitrite and DMSO is, in part, due to a
solvent effect which enhances the apparent value of k6 above
its “proper” value.
HOONO DMSO Me2SO2 MeOH MeSO2H MeSO3H HO•
b
(mM)a
(mM)a
(mM)
(mM)
(mM)
(mM)
(%)c
The measured yields of MSA must be lower than the yields
actually formed in the peroxynitrite/DMSO reaction because
much of the sulfinic acid is oxidized to the sulfonic acid. For
example, the first entry in Table 1 has [peroxynitrite] ) 2.6
mM and [DMSO] ) 105 mM (after mixing) and measured
yields of MSA and MeS(O)2OH of 0.04 and 0.26 mM,
respectively. If we make the assumption that MSA is formed
by hydroxyl radical attack on DMSO (reaction 4) and the
sulfonic acid by further oxidation of the MSA, the percentage
yield of hydroxyl radicals from peroxynitrite is readily calculated
to be 100(0.04 + 0.26)/(2.6 × 0.91) ) 12.5%, where the 0.91
refers to the efficiency of MSA formation in the HO•/DMSO
reaction.30
2.6
2.6
2.6
2.6
2.6
10.5
10.5
2.6d
2.6d
105
105
105
105
105
105
105
420d
420d
0.20
0.21
0.14
0.15
0.18
0.13
0.24
0.36
0.30
0.14
0.16
0.13
0.14
0.16
0.06
0.14
0.17
0.14
0.04
0.10
0.06
0.10
0.12
0.06
0.12
0.19
0.16
0.26
0.21
0.22
0.24
0.18
0.06
0.22
e
12.5
13
12
14
12.5
1.2
3.5
f
e
f
a Concentration after mixing. There was also 100 mM KH2PO4
present after mixing. b Me2SO2 actually formed in the reaction. This
compound is an impurity in the DMSO [(8.1 × 10-2) %]. The measured
concentrations have been corrected by subtracting (8.1 × 10-4)[DMSO]
M. c Calculated as 100([MeSO2H] + [MeSO3H])/([HOONO]0 × 0.91),
see text. d Representative run. e The peak due to MeSO3H was not
resolved from that due to DMSO. f Impossible to calculate.
Interestingly, the extensive conversion of methane sulfinic
acid into methane sulfonic acid cannot simply be attributed to
its oxidation by peroxynitrite (reaction 7) despite the relatively
high rate constant for this reaction (325 M-1 s-1, vide supra).
That is, if the measured concentration of MSA (0.04 mM in
the cited experiment) were to correspond to a steady state
(because it was generated in reaction 4 and consumed by
reaction 7) then the first-order rate constant for peroxynitrite
destruction via reaction 7 would be 4 × 10-5 M × 325 M-1
s-1 ) 0.013 s-1. Such a value is insignificant when compared
with the rate constant for decay of peroxynitrite in the absence
of substrate, ko ) 0.95 s-1. Reaction 7, therefore, plays a
negligible role in the conversion of the sulfinic acid to the
sulfonic acid. Some other oxidizing species must, therefore,
the measured concentration of methyl sulfone in the products
was 0.18 mM. The correction for the sulfone impurity amounts
to 0.08 mM and hence the corrected yield of sulfone is 0.10
mM, or 3.8% based on the initial concentration of peroxynitrite.
If methyl sulfone were the only peroxynitrite/DMSO product
(reaction 6) its relative yield would be given by:
[Me2SO2]/[-OONO(HOONO)] )
k6[DMSO]/(k0 + k6[DMSO])
Taking k0 ) 0.95 s-1 and k6) 2.1 M-1 s-1 as measured in the
stop-flow experiments (vide supra) this equation predicts that
the yield of methyl sulfone should have been 18%. A possible
•
be present in the medium. One reasonable candidate is NO2 ,
reaction 9.

The Complex Chemistry of Peroxynitrite Decomposition
J. Am. Chem. Soc., Vol. 120, No. 29, 1998 7215
MeS(O)OH + NO2 f MeS(O)2OH + •NO
(9)
•
Table 2. Yields of CH3-TEMPOL for the Decomposition of
Peroxynitrite at pH 6.8 and 23 ( 1 ˚C in the Presence of DMSO
and TEMPOL
It should be borne in mind that although the peroxynitrite/DMSO
reaction mixtures were immediately stored at low temperatures
prior to their 1H NMR analyses, further chemical change is not
ruled out simply because all the peroxynitrite had decayed.
Another intriguing product is methanol, which would be most
readily accounted for if the methyl radicals were trapped by
dioxygen, and this was followed by the bimolecular self-reaction
of the resulting methylperoxyl radicals:
HOONO
(mM)a
DMSO
(mM)a
TEMPOL
(mM)a
CH3-TEMPOL
(mM)
HO•
(%)b
2.6
2.6
2.6
2.6
2.6
2.6
2.6
2.6
10.5
10.5
10.5
10.5
42
420
420
420
420
105
105
105
105
420
420
105
105
420
420
105
105
210
53
13
3.3
53
13
3.3
0.8
210
53
210
53
210
53
210
53
0.18
0.19
0.19
0.21
0.17
0.19
0.17
0.15
0.22
0.57
0.21
0.36
0.44
1.15
0.24
0.56
7.7
8.0
8.0
8.8
7.2
8.0
7.2
6.4
2.3
6.0
2.0
3.8
1.2
3.0
0.6
1.5
•
CH3 + O2 f CH3OO•
(10)
(11)
2CH3OO• f CH3OH + CH2O + O2
42
42
42
These two reactions are likely to be the origin of the formal-
dehyde observed in earlier product studies of the peroxynitrite/
DMSO reaction since the solutions of the reactants were not
deaerated.5,32 In contrast, we rigorously excluded oxygen in
our experiments (<10-6 M). However, dioxygen (and nitrite)
are formed during the decomposition of peroxynitrite both in
the presence of hydrogen peroxide12,38-40 and, more importantly,
a Concentration after mixing. There was also 100 mM KH2PO4
present after mixing. b Calculated as 100[CH3-TEMPOL]/([HOONO]0
× 0.91).
10 s after the fast mixing and the solutions were allowed to
stand for 30 or 240 min prior to extraction. Again, no CH3-
TEMPOL could be detected. It is clear, therefore, that CH3-
TEMPOL is a primary product of the peroxynitrite/DMSO
reaction.
also in the absence of any deliberately added hydrogen
39
peroxide.39,40 Thus, Pfeiffer et al.
have reported that, in
addition to nitrate, nitrite and dioxygen (ratio 2:1) are formed
in appreciable yields during the decomposition of H2O2-free
peroxynitrite, e.g., ca. 24% NO2- and 12% O2 at 24 °C and pH
7.4. Roughly similar yields of oxygen were found by Kirsch
et al.40 The nitrite and oxygen yields are independent of the
peroxynitrite concentration (0.01 to 1 mM),39 they increase with
an increase in temperature39 and decrease with a decrease in
pH.39,40 Two different mechanisms for dioxygen and nitrite
formation have been suggested7,39 (see Discussion).
Discussion
Methanol Formation. There can be no doubt that methanol
is formed in the peroxynitrite/DMSO reaction under our
conditions (Table 1). Where does the oxygen come from? The
yield of methanol is far too large for it to be formed by reaction
with residual dioxygen in our nitrogen-purged reactant solutions.
However, dioxygen is known to be formed during the decay of
peroxynitrite, Vide supra, and the reported yields39,40 would
provide sufficient oxygen to account for the formation of
methanol via reactions 10 and 11 in the yields observed in the
present work. Two mechanisms for the formation of oxygen
have been proposed. First,39 by the reaction sequence:
(2) CH3-TEMPOL. The results summarized in Table 2
demonstrate that with 2.6 mM peroxynitrite and either 420 or
105 mM DMSO the yield of CH3-TEMPOL is almost constant
at ca. 8% even when the TEMPOL concentration was reduced
from 210 mM to 3.3 mM, or even to 0.8 mM. Just as was
1
found in the H NMR product analyses, the yield of the HO•
radical-derived product, CH3-TEMPOL, decreased relative to
the initial peroxynitrite concentration as the latter was increased.
Since bicarbonate destroys peroxynitrite20-22 it was no surprise
to find that the deliberate addition of 50 mM bicarbonate to the
buffer reduced the yield of HO• radical-derived products in both
HOONO (HOONO)* f HO• + NO2
(12)
(13)
(14)
(15)
•
HO• + -OONO f HO- + •NO + O2
1
the H NMR and TEMPOL experiments (results not shown).
•
Two sets of additional experiments were conducted to verify
that the CH3-TEMPOL was not being produced via slow
secondary reactions. First, following the usual reaction proce-
dure, the TEMPOL was added in three separate experiments
about 10 s, 30 min and 240 min after the fast mixing of the
peroxynitrite and DMSO solutions. About 20-30 s after these
additions of TEMPOL, the solutions were extracted and
analyzed. No CH3-TEMPOL could be detected. Second,
following the usual procedure, the TEMPOL was added about
NO2 + •NO f N2O3
N2O3 + H2O f 2NO2- + 2H+
However, in our experiments the concentration of DMSO is
very much greater than the concentration of peroxynitrite so
that reaction 13 will not occur and dioxygen would not,
therefore, be formed. Second,7 by the reaction sequence:
HOONO + -OONO a [HOONO/OONO]- (16)
(36) Scaduto, R. C., Jr. Free Radical Biol. Med. 1995, 18, 271-277.
(37) Babbs, C. F.; Gale, M. J. Anal. Biochem. 1987, 163, 67-73. For
related work, see: Babbs, C. F.; Griffin, D. W. Free Radical Biol. Med.
1989, 6, 493-503. Steiner, M. G.; Babbs, C. F. Arch. Biochem. Biophys.
1990, 278, 478-481. Fuku, S.; Hanasaki, Y.; Ogawa, S. J. Chromatogr.
1993, 630, 187-193.
(38) See, e.g.: Alvarez, B.; Denicola, A.; Radi, R. Chem. Res. Toxicol.
1995, 8, 859-864.
(39) Pfeiffer, S.; Gorren, A. C. F.; Schmidt, K.; Werner, E. R.; Hansert,
B.; Bohle D. S.; Mayer, B. J. Biol. Chem. 1997, 272, 3465-3470.
(40) Kirsch, M.; Lomonosova, E. E.; Korth, H.-G.; Sustmann, R.; de
Groot, H. J. Biol. Chem. In press, and private communication.
[HOONO/OONO]- f 2NO2- + H+ + O2
(17)
However, peroxynitrite decay gives mainly nitric acid via a
kinetically first-order process (at least at relatively low con-
centration of peroxynitrite (2.6 mM)) and so this second
mechanism would seem to imply that at a given pH the yield
of oxygen would increase as the concentration of peroxynitrite
increases. This appears not to be the case from 0.01 to 1 mM

7216 J. Am. Chem. Soc., Vol. 120, No. 29, 1998
Richeson et al.
peroxynitrite.39 Thus, neither of the proposed mechanisms for
oxygen formation appear able to account for methanol formation
via reactions 10 and 11 under our experimental conditions. Of
course, if oxygen is actually formed by some third mechanism,
then methanol (and formaldehyde) can readily be accounted for
by these two reactions. However, there are other potential
oxidants of the methyl radical which could yield methanol, of
Earlier claims for much lower hydroxyl radical yields (0-
4%, see Introduction) probably arose from three main prob-
lems: (i) adventitious bicarbonate,20-22 (ii) the use of high
concentrations of peroxynitrite,22,34 and (iii) direct, bimolecular
reactions of peroxynitrite with the probe molecule or its expected
hydroxyl radical-induced product. Unfortunately, the possibility
of such direct reactions was not always explored. For example,
low hydroxyl radical yields based on spin trapping (0.9%,13
zero,15 and probably zero16-18) would appear to be unreliable
because Lemercier et al.16 have shown that at least one of the
spin traps employed (DMPO) reacts directly with peroxynitrite
in a bimolecular reaction.45 Other experimental results which
have been used to argue against the simple homolysis of
•
which the most likely would appear to be NO2 and, less
•
probably, the peroxynitrite itself (assuming that CH3 is not
oxidized to CH3+). We, therefore, suggest that methanol may
be formed by reaction 18 and/or 19.
CH3 + NO2 f CH3ONO H O8 CH3OH + NO2- + H+
2
•
•
•
peroxynitrite to HO• and NO2 radicals have included the lack
(18)
of effect of viscosity on the rate of peroxynitrite decomposi-
tion.19 However, in our opinion, it is doubtful that these
measurements would actually be capable of clearly detecting a
yield of free HO• radicals of only ca. 10%. We hold the same
doubts about Kissner et al.’s7 report that favored a nonhomolytic
isomerization of peroxynitrite to nitrate on the basis of a volume
of activation for peroxynitrite decomposition which is consider-
ably smaller than that for a “normal” bond homolysis, viz., 1.7
cm3 mol-1 vs 10 cm3 mol-1. However, these authors were
careful to state the following: “the observed activation volume
may be compatible with (i) a lengthening of the O-O bond
followed by isomerization..., (ii) isomerization mainly as
described under (i) with a small fraction isomerizing Via
homolysis (our italics), and, possibly, (iii) heterolysis into
hydroxide and nitryl, in which case electrostriction would lower
the activation volume of the heterolysis,” vide infra. We
therefore suggest that our experimental evidence for a ca. 10%
yield of HO• should be accepted. No doubt future work will
refine this value and, we are confident, will confirm our
assumption that peroxynitrite’s powerful oxidant is, indeed, the
HO• radical and not some ill-defined “hydroxyl radical-like
species”.
CH3 + HOONO (-OONO + H+) f CH3OH + NO2 (19)
•
•
The Yield of Hydroxyl Radicals from Peroxynitrite. We
will make the assumption that MSA and methyl radicals can
only be formed from DMSO by its reaction with free hydroxyl
radicals. We will also assume that only the data obtained at
our lowest peroxynitrite concentration, 2.6 mM, are meaningful
because the kinetic complexities in the decay of peroxynitrite
at elevated concentrations, vide supra,7,34 would appear to
decrease the efficiency of HO• radical formation. With these
two assumptions, our quantitative analyses of the products of
the peroxynitrite/DMSO reaction indicate that the thermal
decomposition of peroxynitrite in water at ambient temperatures
gives free hydroxyl radicals in yields of about 13% (1H NMR
analyses, Table 1) or about 8% (CH3-TEMPOL analyses, Table
2). We do not understand why our two semiindependent
approaches do not yield analytical data which are in better
agreement. However, both sets of experimental data imply that
free HO• radicals are produced during the decomposition of 2.6
mM peroxynitrite in a significant yield that is probably best
given as 10 ( 3%. We note with particular pleasure that this
hydroxyl radical yield of 10 ( 3% is in excellent agreement
with the 7% yield of malonaldehyde produced from deoxyribose
reported by Beckman et al.5 in their 1990 landmark paper and
with the 8% yield of formaldehyde produced from DMSO
reported by Pryor et al.32
We cannot explain all earlier claims for HO• radical yields
from peroxynitrite which are very much larger than 10% (see
Introduction). However, a referee has pointed out that Crow
et al.’s11 calculated yield of HO• radicals (25%) was based on
the oxidation of 2,2′-azinobis(3-ethyl-1,2-dihydrobenzothiazoline
Our confidence in our conclusion that ca. 10% free HO•
radicals are produced is bolstered by two other pieces of
evidence. First, Beckman et al.11 have shown that, in the
•
6-sulfonate) (ABTS) to its radical cation and that both NO2
(formed in 16% yield)11 and HO• (10 ( 3% yield) can perform
this one-electron oxidation to give a combined oxidizing effect
essentially equal to the reported 25%.
•
presence of 50 mM DMSO, NO2 is formed in a yield of ca.
16% at pH 6.8.43 Second, Butler et al.44 have demonstrated,
using 15N CIDNP, that tyrosine nitration by 15N-labeled per-
oxynitrite at pH 12 yields 15N NMR signals from nitrate in
enhanced absorption, which indicates a reaction of the caged
We tentatively attribute the low efficiency of free radical
formation by peroxynitrite to a rather large cage effect induced
by hydrogen bonding of peroxynitrous acid and its fragments
by the surrounding water molecules. This idea was first
considered seriously by Pryor and Squadrito.6 These authors
also pointed out that, in addition to the “stickiness” of the water
molecules (caused by hydrogen bonding) that surround the HO•/
HO•/NO2 radical pair, and from nitrite in emission, which
•
indicates that nitrite is formed from radicals which escape the
cage, i.e., free radicals. Unfortunately, these CIDNP experi-
ments were done without exclusion of CO2 so they are not
definitive evidence for the formation of radicals from perox-
ynitrite under CO2-free conditions.
•
NO2 singlet geminate radical pair, there could also be a
pronounced polar effect that would further enhance cage return,
i.e.,
(41) A much larger volume of activation (9.6 ( 1.0 cm3 mol-1) has
also been reported.42 However, the experimental work was much more
limited, including only one pressure different from atmospheric (vs 8),7
one temperature (vs 2),7 and one pH (vs 2).7 The validity of Goldstein et
al.’s42 large volume of activation therefore remains uncertain.
(42) Goldstein, S.; Meyerstein, D.; van Eldik, R.; Czapski, G. J. Phys.
Chem. A 1997, 101, 7114-7118.
+
However, the HO-/NO2 ion pair would appear to be more
stable in water than the HO•/NO2• radical pair by about 9 kcal/
•
(43) An earlier publication gives a somewhat higher yield of NO2 , see:
Zhu, L.; Gunn, C.; Beckman, J. S. Arch. Biochem. Biophys. 1992, 298,
452-457.
(44) Butler, A. R.; Rutherford, T. J.; Short, D. M.; Ridd, J. H. J. Chem.
Soc., Chem. Commun. 1997, 669-670.
(45) The unreliability of HO• yields based on spin trapping is further
emphasized by the fact that the HO•/DMPO spin adduct had already been
shown to react directly with and be destroyed by peroxynitrite.46 Other spin
traps have also been shown to react directly with peroxynitrite.17.

The Complex Chemistry of Peroxynitrite Decomposition
J. Am. Chem. Soc., Vol. 120, No. 29, 1998 7217
mol.47 This would make it difficult for the ion pair to generate
the radical pair. Perhaps the ion pair is formed only slowly
from the radical pair because this electron transfer reaction
would involve a large solvent reorganization energy. A slow
formation of the ion pair would allow time for some radicals to
escape the solvent cage. Once formed, however, the ion pairs
probably collapse to produce mainly (or solely) nitric acid.
The geminate radical pairs may be held in the positions in
which they are produced by the surrounding water molecules
so that most in-cage reactions yield peroxynitrite with only a
small fraction of the radical pairs reorienting in-cage and
collapsing to nitric acid.48 (As Pryor and Squadrito6 noted, a
high ratio of in-cage return to give peroxynitrite relative to
rearrangement to nitrite could explain the relatively low
Arrhenius preexponential factor for peroxynitrite decomposition,
vide infra.) Thus, our results suggest that ca. 90% of the
peroxynitrite isomerizes to nitric acid by a rather inefficient in-
in water? Gas-phase thermokinetic properties imply that the
Arrhenius preexponential factor for unimolecular O-O bond
cleavage of HOONO (kf of reaction 3) should be ca. 6 × 1015
s-1, which leads to a gas-phase rate constant for O-O bond
homolysis of ca. 500 s-1 (see Appendix A). This is 500 times
greater than the experimental rate constant of ca. 1 s-1 measured
for peroxynitrite decay in water. Provided the activation
enthalpy is unchanged at 18 kcal mol-1, this means that the
Arrhenius preexponential factor must be 500 times lower in
water than in the gas phase, i.e., ca. 1013 s-1 62
The simplest
.
explanation for this unexpectedly low A factor is that only 1
out of every 500 O-O bond homolysis events leads to products.
That is, of every 500 geminate singlet HO•/NO2 radical pairs
•
created in water there are 499 which undergo cage return to
reform HOONO and only one which forms products and hence
leads to a measurable loss of HOONO. Our results indicate
that this 1 in 500 event yields nitrate and free HO• (and NO2 )
•
cage reaction of the geminate HO•/NO2 radical pair with
•
radicals in a roughly 9:1 ratio. However, we do not claim that
HO• radical production from peroxynitrite is necessarily
important in vivo. In the first place, the very formation of
peroxynitrite in vivo has been seriously questioned65 as has its
role as a cytotoxic agent in the immune response.65 Second, if
peroxynitrite is, indeed, formed in vivo then some fraction of
it will undoubtedly undergo direct, bimolecular reactions with
the many available biological target molecules (e.g., thiols,
ascorbate, and, particularly,20,21,61 bicarbonate) rather than
isomerizing to nitrate with a concomitant release of HO• and
+
possibly some minor contribution from the geminate HO-/NO2
ion pair.50 The remaining ca. 10% of the geminate radical pairs
•
yield freely diffusing HO• and NO2 radicals. The products
derived from these free radicals depend on the presence and
relative concentrations of compounds capable of reacting with
them. Specifically, for the HO• radical these may include
-OONO 39 and NO2- 49b but, in our experiments, the main sink
for HO• is DMSO (because of its high concentration relative to
the two inorganic ions).
In support of the above picture we note that the measured
activation enthalpy for peroxynitrite decomposition, viz.,24 18
( 1 kcal mol-1, lies in the range of calculated O-O bond
dissociation energies (BDEs) which have been reported for
HOONO in the gas phase, viz.54 16-23 kcal mol-1. Our own
thermochemical arguments (see Appendix A) also lead to BDE-
[HO-ONO] ) 18 ( 1 kcal mol-1. An O-O BDE of 18 kcal
mol-1 raises an intriguing question that others have wrestled
with,6,19,23,24,60,61 viz., why is peroxynitrous acid so long-lived
NO2 radicals. Only future work will determine whether HO•
•
radical formation from peroxynitrite in vivo is biologically (and
medically) significant.
Experimental Section
Materials. Compounds were the purest available commercially and
were used as received: sodium azide, TEMPOL, methyl sulfone
(Aldrich); DMSO (BDH Inc.); and methane sulfinic acid (Lancaster
Synthesis Inc.).
Synthesis of Peroxynitrite. The procedure of Gleu and Roell31 and
Pryor et al.32 was employed. To avoid any contamination with (bi)-
carbonate from ambient air, the handling of the reagents and the
synthesis of the peroxynitrite were performed inside a glovebag under
a nitrogen atmosphere. Air inside the bag was removed by multiple
flushings with nitrogen and the bag was then kept at slightly above
(46) Augusto, O.; Gatti, R. M.; Radi, R. Arch. Biochem. Biophys. 1994,
310, 118-125. Augusto, O.; Radi, R.; Gatti, R. M.; Va´squez-Vivar, J.
Methods Enzymol. 1996, 269, 346-353.
(47) Stanbury, D. M. AdV. Inorg. Chem. 1987, 33, 69-138. We are
indebted to an unknown referee for bringing this important fact to our
attention.
(48) It has been shown by pulse radiolysis that even free HO• radicals
•
can react with NO2 to form peroxynitrite by monitoring this reaction at
(55) McGrath, M. P.; Francl, M. M.; Rowland, F. S.; Hehre, W. J. J.
Phys. Chem. 1988, 92, 5352-5357.
302 nm49 where the peroxynitrite absorbs. Nitrate formation was not
monitored and therefore the ratio of the yields of peroxynitrite to nitrate
(56) Calculation at the B3LYP/6-311++G(d, p) level gives 12.3 kcal
•
from freely diffusing HO• and NO2 radicals remains to be established.
mol-1 57a
. However, calculations at such a high level underestimate O-O
However, if this ratio is large (as is probably the case: private communica-
tion from Professor W. H. Koppenol) then there is no longer any need to
bond strengths by 6-10 kcal mol-1 and O-H BDE’s by 3-7 kcal mol-1 57
.
Calculations at the lower B3LYP/6-31G* level give 20 kcal mol-1 57a.
.
•
invoke a slow reorientation of the geminate HO•/NO2 radical pair due to
(57) (a) Korth, H.-G. Universita¨t-GH Essen. Unpublished results. (b) van
Scheppingen, W.; Dorrestijn, E.; Arends, I.; Mulder, P.; Korth, H.-G. J.
Phys. Chem. A 1997, 101, 5404-5411.
hydrogen bonding from the surrounding water molecules to explain the
large extent of in-cage return to peroxynitrite.
(49) (a) Gra¨tzel, M.; Henglein, A.; Taniguchi, S. Ber. Bunsen-Ges. Phys.
Chem. 1970, 74, 292-298. (b) Logager, T.; Sehested, K. J. Phys. Chem.
1993, 97, 6664-6669.
(58) Burkholder, J. B.; Hammer, P. D.; Howard, C. J. J. Phys. Chem.
1987, 91, 2136-2144.
(59) Bach, R. D.; Ayala, P. Y.; Schlegel, H. B. J. Am. Chem. Soc. 1996,
118, 12758-12765.
(50) A minor contribution to nitric acid formation from the ion pair is
consistent with a recent report of 13% 18O incorporation into the nitrate
product when peroxynitrite was decomposed in CO2-free H218O (pH 6.8)
while 83% of the nitrate showed no 18O incorporation.51,52 That is, part of
(60) Tsai, H.-H.; Hamilton, T. P.; Tsai, J.-H. M.; van der Woerd, M.;
Harrison, J.-G.; Jablonsky, M. J.; Beckman, J. S.; Koppenol. W. H. J. Phys.
Chem. 1996, 100, 15087-15095.
+
the 18O-containing nitrate arises from the NO2 + H218O reaction and the
(61) Houk, K. N.; Condroski, K. R.; Pryor, W. A. J. Am. Chem. Soc.
1997, 119, 2964. Correction: 1997, 119, 2964.
remainder from reaction of free NO2• radicals with the H218O. Unfortunately,
the experimental technique (Raman spectroscopy) did not allow determi-
nation of the 18O content of the nitrite product.
(62) Such a large reduction of the Arrhenius A factor for O-O bond
homolysis on changing from the gas phase to a solvent that interacts with
the product oxyl radicals is not without precedent. For O-O homolysis of
(51) Bohle, D. S.; Hansert, B. NITRIC OXIDE: Biol. Chem. 1997, 1,
502-506.
(52) There is an earlier report that when peroxynitrite rearranges to nitrate
none of the oxygen atoms come from water.53 We are indebted to Professor
W. A. Pryor for pointing out that this result is consistent with the “sticky
cage” concept.
di-tert-butyl peroxide the A factors are 7 × 1015 s-1 in the gas phase,63
CH3CN.64
4
× 1015 s-1 in THF,64 5 × 1014 s-1 in benzene,64 but only 4 × 1012 s-1 in
(63) Lossing, F.; Tickner, A. W. J. Chem. Phys. 1952, 20, 907-914.
(64) Calculated from data given in the following: Reichardt, C. SolVents
and SolVent Effects in Organic Chemistry; VCH: New York, 1988, Tables
5-14, p 187.
(53) Anbar, M.; Taube, H. J. Am. Chem. Soc. 1954, 76, 6243-6247.
(54) Individual BDE values are 16,55 16-18,56 20,58 21,18 22,59 and 2360
kcal mol-1, with errors generally estimated as (2-3 kcal mol-1
.
(65) Fukuto, J. M.; Ignarro, L. J. Acc. Chem. Res. 1997, 30, 149-152.

7218 J. Am. Chem. Soc., Vol. 120, No. 29, 1998
Richeson et al.
atmospheric pressure at all times. Doubly distilled water (100 mL)
was purged with nitrogen, placed in the bag, and purged for a further
30 min. Sodium hydroxide (1 pellet, ca. 0.1 g, from a new container
that was only opened in the bag) was added to the water, followed by
sodium azide (1.3 g, 200 mM, taking the same precautions). This
solution (pH ∼12) was poured into a nitrogen-containing 3-arm flask
that was then sealed with septa, removed from the bag, placed in an
ice bath and flushed with nitrogen through a hypodermic needle. A
glass frit was immersed in the solution through which ozone was
bubbled for 100 min. The ozone was made by passing oxygen (dried
and passed over NaOH pellets) at a rate of ca. 0.3 mL/min through a
Weisbach Laboratory Ozonator (Model T-816). Finally, the reaction
mixture was again purged with nitrogen for more than 30 min to remove
dissolved oxygen. The vessel ports were then sealed with septa and
the flask was returned to the glovebag. The solution was divided into
5 mL sample vials which were then sealed, removed from the bag,
and stored at -80 °C. The peroxynitrite concentration was calculated
was either used directly or used after the addition of a known quantity
of a substrate in the glovebag. Both solutions were then purged with
nitrogen for 30 min before being drawn into gas tight syringes. The
entire stop-flow apparatus was flushed with nitrogen for 30 min prior
to each experiment. The syringes containing the two solutions were
removed from the bag and attached to the stop-flow apparatus. Equal
volumes of the two solutions were rapidly mixed (∼1.3 ms) and the
decrease in absorbance at 302 nm was monitored. Reactions occurred
at pH 6.8 (determined after each experiment) and at room temperature
(23 ( 1 °C, automatically recorded).
Products. The products of the reaction of peroxynitrite with DMSO
alone in the KH2PO4 buffer and with DMSO plus TEMPOL in the
buffer were generated at room temperature (23 ( 1 °C) as in the above-
described kinetic experiments but with a rapid mixing (∼ 4 ms)
apparatus with no kinetic capabilities. A glass collection vessel (ca. 2
mL volume) was attached to the mixer. The collection vessel had
stopcocks on both ends. It was pre-flushed with the mixed reactants
before the products were collected for analyses.
from the absorbance at 302 nm (ꢀ ) 1670 M-1 cm-1 66 after a 30-fold
)
dilution with 0.1 N NaOH.32 Samples were withdrawn from the 5 mL
vials after thawing in the glovebag by using a hypodermic syringe
within the bag.
In preliminary (monitoring) experiments we found that our concen-
tration vs time plots for azide consumption (method of Herzog and
Rudolf67 with a Nicolet 20DXB FTIR spectrometer) and peroxynitrite
formation were (surprisingly) almost indistinguishable from those shown
in Figure 2 in Pryor et al’s publication.32 Although azide was below
our detection limits in the final product is was present in sufficient
concentration to make the determination of MSA yields by reaction
with diazonium salts quite impossible.35
Preparation of 200 mM KH2PO4 Buffer. This solution was also
made up in the nitrogen-filled glovebag. The correct amount of KH2-
PO4 was weighed out, transferred to the bag, and, after some time,
added to a flask containing nitrogen-purged, doubly distilled water.
Chelex was then added and the flask was sealed with a septum and
stirred overnight in the bag. Samples of buffer were withdrawn in the
glovebag with a hypodermic syringe.
The collected products from the peroxynitrite/DMSO reactions were
1
diluted with 15% (v/v) D2O and analyzed by H NMR (400 MHz).
The following compounds were identified and quantified relative to
the known concentration of DMSO (2.61 ppm): CH3OH (3.24 ppm),
(CH3)2SO2 (3.03 ppm), CH3SO3H (2.70 ppm) and CH3SO2H (MSA,
2.19 ppm).
For the peroxynitrite/(DMSO + TEMPOL) reactions, a 1 mL aliquot
was removed from each collected product sample within 1 min of its
generation. To this aliquot was added a known amount of CD3-
TEMPOL to serve as an internal standard. The aliquot was then
extracted five times with ether (5 × 1 mL), the ether extract was dried,
and the ether was removed by using a rotovap. After addition of a
small quantity of ether, the yield of CH3-TEMPOL (parent ion m/e )
187) was determined quantitatively relative to the CD3-TEMPOL (m/e
) 190) internal standard on a Hewlett-Packard 5790 Series gas
chromatograph with a 5970 Series mass selective detector operated in
the single ion monitoring (SIM) mode. Blank experiments with equal
concentrations of CH3-TEMPOL and CD3-TEMPOL gave m/e peaks
at 187 and 190 of equal intensity.
Synthesis of CH3-TEMPOL and CD3-Tempol. Surprisingly, no
physical properties have been reported for either of these hydroxyl-
amines.68 Into an oven-dried 50-mL three-necked flask was placed
0.2 g (1.16 mmol) of TEMPOL and 15 mL of anhydrous ether. The
resulting solution was cooled to -78 °C and stirred under nitrogen,
and methyllithium (1.66 mL of a 1.4 M solution in ether, 2.32 mmol)
was added in one portion with a syringe. The reaction mixture was
slowly warmed to room temperature and was held at this temperature
for 1 h, during which time the solution turned pale orange. After the
mixture was cooled to 0 °C, 5 mL of distilled water was added dropwise
with continuous stirring. The aqueous phase was extracted with ether
(3 × 10 mL) and the organic phases were combined, dried with MgSO4,
filtered, and concentrated by rotovap to yield 0.14 g (66%) of crude
CH3-TEMPOL. This was purified by column chromatography with
ether as eluent and recrystallized from n-pentane at -78 °C: white
Appendix A
The reaction enthalpy (∆20H) and reaction entropy (∆20S) at
298 K can be derived as follows:
HOONO {kkf
\
} HO• + NO2 f H+ + NO3
(20)
•
-
r
With a basis set containing ∆fH° and S° for CH3OH, CH3-
OOH, CH3ONO, HONO, and H2O2,69 the heat of formation of
HOONO is computed according to group increment rules
(replacing H by NO or H by OH) as -0.7 and -2 kcal mol-1
:
mean value, ∆fH°(HOONO) ) -1 ( 1 kcal mol-1. Accord-
ingly, the entropy gain for H f NO amounts to 10.7 cal mol-1
K-1 and that for H f OH amounts to 10.2 cal mol-1 K-1: thus
the mean S°(HOONO) ) 68 ( 1 cal mol-1 K-1. Combination
with OH and NO2 leads to the thermodynamic values for
equilibrium 20 at 298 K of ∆20H° ) 18 ( 1 kcal mol-1, ∆20S°
) 33 ( 1 cal mol-1 K-1, and ∆20G° ) 8 ( 1 kcal mol-1, which
is in excellent agreement with Koppenol and Kissner’s27 value
1
crystals, mp 90-91 °C; H NMR (300 MHz, CDCl3, ppm) 3.94 (m,
1H), 3.60 (s, 3H), 1.79 (dd, J ) 11.6, 3.0 Hz, 2H), 1.45 (dd, J ) 11.7,
11.6 Hz, 2H), 1.26 (d, J ) 7.0 Hz, OH), 1.21 (s, 6H), 1.13 (s, 6H).
The same procedure with CD3Li yielded CD3-TEMPOL: white
1
crystals, mp 91-92 °C; H NMR (300 MHz, CDCl3, ppm) 3.94 (m,
1H), 1.79 (dd, J ) 11.7, 3.0 Hz, 2H), 1.46 (dd, J ) 11.7, 11.3 Hz,
2H), 1.25 (d, J ) 7.0 Hz, OH), 1.21 (s, 6H), 1.13 (s, 6H).
Kinetics. The kinetics of peroxynitrite decay were monitored at
302 nm with a Biosequential SX-18MV stopped-flow reaction analyzer
(Applied Photophysics). The stock peroxynitrite solution was, if
needed, diluted with nitrogen-purged, doubly distilled water made to
pH 12 in the glovebag as described above. The KH2PO4 buffer solution
for ∆20G° of 7.2 kcal mol-1
.
In a large number of studies pertaining to the gas phase70 the
rate constant for combination has been established to be kr )
k-20 ) 2.4 × 1010 M-1 s-1. This is an average, high-pressure
limit, value and is independent of the temperature. Combination
of kr ()k-20) with the thermodynamic data given above yields
(66) Hughes, M. N.; Nicklin, H. G. J. Chem. Soc. A 1968, 450-452.
(67) Herzog, K.; Rudolph, W. Fresenius, J. Anal. Chem. 1992, 343, 619-
620.
(68) CH3-TEMPOL has been made previously but its physical and
spectroscopic properties have not been reported. See: Sheats, J. R.;
McConnell, H. M. J. Am Chem. Soc. 1977, 99, 7091-7092. Bradley, P.;
Suardi, G.; Zipp, A. P.; Eisenberg, R. J. Am. Chem. Soc. 1994, 116, 2859-
2868.
(69) Stein, S. E.; Rukkers, J. M.; Brown, R. L. NIST Structures and
Properties Database, version 2.0; NIST Standard Reference Data; National
Institute of Standards and Technology; Gaithersburg, MD, 1994.
(70) Tsang, W.; Herron, J. T. J. Phys. Chem. Ref. Data 1991, 20, 609-
663.

The Complex Chemistry of Peroxynitrite Decomposition
J. Am. Chem. Soc., Vol. 120, No. 29, 1998 7219
the preexponential (A20) at 298 K as follows:
regarding many aspects of peroxynitrite chemistry. We also very
much appreciate the many insightful comments and useful
suggestions made by Dr. Hans-Gert Korth, Professor Willem
H. Koppenol, and some unknown referees. One of us (K.U.I.)
would like to thank Professor J. S. Beckman for stimulating
his interest in, and encouraging us to work on, that fascinatingly
“simple” compound, peroxynitrite. We also thank Drs. J.
Lusztyk, J. T. Banks, P. A. MacFaul, and T. Paul for much
advice and encouragement, Sue Lin for the synthesis of
analytical samples of CH3- and CD3-TEMPOL, Don Leek and
A20/s-1 ) A-20(0.229T)-1 exp(∆20Sq/R)
) (2.4 × 1010)(1.5 × 10-2)(1.64 × 107) )
5.9 × 1015
Thus, with ∆20Hq) ∆20H ) 18 kcal mol-1 and ∆20Sq ) 11.6
cal mol K-1 (from A20 ) (5.66 × 1010)T exp(∆20SqR)), the rate
constant for homolysis of HOONO in the gas phase becomes
kf()k20) ) (5.9 × 1015) exp(-18/RT) s-1 ) 465 s-1 at 298 K
(where R ) 1.987 × 10-3 kcal mol-1).
1
Leanne Doiron for the H NMR spectra, and the Association
for International Cancer Research and the National Foundation
for Cancer Research for partial support of this work.
Supporting Information Available: Kinetic data for the
reactions of peroxynitrite with DMSO, MSA, and TEMPOL
(Tables 3-5, (3 pages, print/PDF). See any current masthead
page for ordering information and Web access instructions.
Acknowledgment. This paper is dedicated to Professor
William A. (Bill) Pryor, chemist and pioneer in chemical
research on the role of free radicals and other oxidants (including
peroxynitrite) in health and disease. We are deeply indebted to
Professor Pryor for his very helpful and generous advice
JA980871X